Rhizobium rhizogenes‐mediated hairy root induction and plant regeneration for bioengineering citrus

Plant transformation is widely used in functional genetics and bioengineering (BE)-based crop improvement. With the advent and deployment of CRISPR-based genome editing technology and increased public awareness of the safety of BE crops, the applications of plant genetic engineering are bound to increase. Multiple methods have been used for plant transformation, such as DNA particle bombardment, PEG-mediated transfection, and Agrobacterium-mediated transformation. Although these tools were used to transform several crops, the efficiency of genetic modification of tree crops such as Citrus spp. has been very challenging. This is mainly due to the hardy nature of the tissues, slow growth, and recalcitrance to Agrobacterium tumefaciens infection (Dominguez et al., 2022; Pena et al., 1995). Despite improvements, the widely used A. tumefaciens- mediated transformation in citrus has low transformation efficiencies (~0.1%–11%) depending on the cultivar and can take 6–12 months to recover transgenic plants (Dominguez et al., 2022; Pena et al., 1995). Further multiplication of the citrus plants to generate enough clones often would require additional time (~3–6 months; Dominguez et al., 2022; Pena et al., 1995). Plant cells have high levels of developmental plasticity and can dedifferentiate to new cell types under appropriate environmental and growth conditions (Grafi, 2004). Transgenic "hairy" roots and stable transgenic plants can be induced using Rhizobium rhizogenes (previously Agrobacterium rhizogenes). Like A. tumefaciens, R. rhizogenes introduces its root-inducing transfer DNA (Ri T-DNA), which encodes multiple root locus (rol) genes into plant cells. The expression of rol genes in transformed shoot cells is presumed to promote their dedifferentiation into root cells. The resulting transgenic roots are anatomically and metabolically similar to normal roots (Ron et al., 2014) and have been widely used in plant functional biology, biotechnology, metabolic engineering, molecular pharming, and studies of root–rhizosphere interactions (Cui et al., 2020; Ono and Tian, 2011; Ron et al., 2014). However, the various applications of R. rhizogenes for recalcitrant citrus are limited. We recently showed the utility of R. rhizogenes to induce transgenic hairy roots for expressing transgenes, CRISPR-constructs, and screening antimicrobials effective against fastidious pathogens in citrus and potato (Irigoyen et al., 2020). Here, we show that R. rhizogenes can be used to induce transgenic hairy roots from diverse citrus cultivars such as grapefruit (Rio Red, Citrus × paradisi Macf), sweet orange (Valencia, Citrus × sinensis), rough lemon (Citrus × jambhiri Lush), carrizo (Citrus × poncirus), and citron (Citrus medica) at efficiencies of 28%–75% (Figure S1). Furthermore, as a proof-of-concept, we use carrizo to demonstrate the recovery of stable transgenic plants using transgenic hairy roots (Figure 1). Inducing transgenic hairy roots and plant regeneration is comprised of two main phases. In Phase I, R. rhizogenes transformed citrus (carrizo) epicotyls in vitro (Figure 1a–e). Briefly, explants were immersed in a suspension of R. rhizogenes (OD600 = 1.0) carrying a binary vector, followed by co-cultivation for 3 days. Explants were subsequently transferred to selection media with appropriate antibiotics. Explants typically produced ~2–10 independent roots, which were further subjected to molecular diagnostics and quality control (QC; Figure 1f–i). Depending on the cultivar, citrus can naturally produce non-transgenic adventitious roots tolerating various antibiotic selections. Hence, a fluorescence reporter (e.g., GFP) based visual screening of the newly emerging roots was helpful in screening and selecting transgenic roots (Figure 1g). In addition, PCR-based diagnostics were performed on independent transgenic roots to confirm the presence of the transgene (e.g., GFP) in the transformed roots and eliminated R. rhizogenes contamination (Figure 1h and Figure S2). As a proof-of-concept for CRISPR-editing applications in citrus roots, an endogenous gene (CsNPR3) was edited. Confirmation for CsNPR3-editing was performed by amplicon sequencing (Figure 1i). These QC steps were vital to validating and selecting the desired transgenic roots for subsequent experiments. Furthermore, as an alternative to the in vitro approach, hairy root induction can also be achieved ex vivo using greenhouse-grown citrus shoots as a source for transformation (see Supplementary methods). However, the transgenic roots must be thoroughly surface sterilized before reintroducing them to the next regeneration steps in vitro. Validated carrizo transgenic roots were subsequently advanced to Phase II, that is, regeneration of transgenic plants and clonal propagation (Figure 1j–m). Each independent transgenic root was sectioned into ~0.5–1 cm segments and placed on shoot regeneration media (Figure 1j). Multiple media were tested for optimal carrizo shoot regeneration by modifying growth hormones (e.g. IAA, IBA, BAP, GA3, Zeatin; Table S1 and Figures S3 and S4). RSRM1 and RSRM3 with either IAA/IBA (1 mg/L) or BAP (1 mg/L) regenerated normal-looking shoots from carrizo roots, however. RSRM2 with Zeatin (2.2 mg/L), IAA (0.01 mg/L), and GA3 (0.2 mg/L) produced abnormal shoots (Figure S3). For recalcitrant cultivars, such as Rio Red Grapefruit, Mexican Lime, Swingle Citrumelo, and Valencia Sweet Orange, higher levels of BAP (RSRM3.2 and RSRM3.3) produced normal-looking shoots from roots (Table S1). Typical progression of citrus shoot regeneration from roots occurs via an intermediate dedifferentiated callus formation that occurs ~2 weeks in the regeneration media (Figure 1k). The callus subsequently differentiated into shoots, ready for excision and rooting by ~4–6 weeks (Figure 1l). An independent root can produce multiple individual shoots (Figure 1l). Because the shoots were derived from the same confirmed transgenic root tissue, they are all clones, thus eliminating the need for additional lengthy multiplication or bulking steps. All excised shoots were maintained in the rooting media, where they would grow further for ~6 weeks or when they were robust enough to move to the soil (Figure 1m). The root-derived plants retained stable expression of the transgene (e.g., GFP encoded on the T-DNA) (Figure S5H) and grew normally (Figure 1l,m). Interestingly, expression of root locus (rolB and rolC) genes, encoded on the Ri-TDNA and co-transferred into the plant genome (Figure S5B,C), was undetectable in transgenic shoots (Figure S5F,G). Possible explanations could be stochastic transcriptional or post-transcriptional gene silencing of rol locus or poor rol promoter activity in citrus shoots. Regardless, the lack of rol expression/activity could be congenial for shoot regeneration and also explains the normal growth phenotypes of the recovered transgenic plants (Figure 1l,m). The proposed R. rhizogenes-mediated citrus hairy root induction, shoot regeneration, and multiplication process can be achieved in ~6 months vs. ~12–18 months typically spent in A. tumefaciens-based transformation. In addition, the R. rhizogenes-mediated root induction is highly efficient. We evaluated the competence and hairy root induction frequency of six different citrus cultivars, including multiple scions (grapefruit, sweet orange), rootstocks (carrizo, sour orange, rough lemon), and a true citrus species (Citrus medica, a.k.a. citron) (Figure S1). The hairy root induction frequency was estimated using the formula: Number of GFP-positive roots/total number of roots obtained ×100. The frequency ranged from 25% to 75% among the citrus cultivars and was comparable to R. rhizogenes-mediated transformation efficiencies reported previously for citrus (Ma et al., 2022; Perez-Molphe-Balch and Ochoa-Alejo, 1998; Xiao et al., 2014) and other recalcitrant dicot plants (Cui et al., 2020). The frequency of sweet orange and grapefruit ranged from ~37% to 75% (Figure S1). The frequency among rootstocks (carrizo, rough lemon, and sour orange) ranged from ~65% to 75% (Figure S1). The true citrus species, Citrus medica (citron), was a prolific rooter and produced roots in as little as ~2–3 weeks, but ~70% were non-transgenic (adventitious) and reduced the transgenic root formation frequency to ~28% (Figure S1). Overall, ~25–75% hairy root induction efficiency was attained in diverse citrus cultivars using R. rhizogenes-mediated transformation. Furthermore, generating multiple clonal plants using root tissues is helpful for the micropropagation of citrus (Figure 1l). Interestingly, we observed differences in shoot regeneration potential along the root length in carrizo. Better shoot regeneration occurred from root sections away from the root apex, where auxin levels are the lowest (Figure 1j–l). We speculate that the physiological auxin gradients along the root axis, with the maximum levels at the root apex, could be inhibitory for shoot differentiation. Nevertheless, multiple shoots and plants can be recovered from a single root. In conclusion, we demonstrate a versatile R. rhizogenes-mediated hairy root induction, plant regeneration, and clonal propagation approach that could be helpful for multiple bioengineering and gene-editing applications in citrus and other recalcitrant tree crops. The authors thank Shreya Udawant, Anissa Solis, Snehalatha Parida, Corinne Laughlin (Texas A&M AgriLife Research), and TAMU-Kingsville Citrus Center for various citrus resources and technical assistance. This study was partly supported by funds from USDA-NIFA (2021-70029-36056; HATCH TEX09621), FFAR (21010111), Texas A&M AgriLife Research Insect-vectored Disease Seed Grants (114185-96210), and the Texas A&M AgriLife Institute for Advancing Health Through Agriculture to KM. All authors declare no competing interests. K.M. conceived, designed, and supervised the experiments. M.R., M.D., S.I., and C.P. designed, performed the experiments and analysed the data. All authors contributed to the preparation and editing of the final manuscript. Table S1 Media used for citrus hairy root induction and root-to-shoot regeneration. Figure S1 Rhizobium rhizogenes-mediated induction of transgenic roots from different citrus cultivars. Figure S2 Molecular diagnostics of transgenic hairy roots. Figure S3 Effects of growth media on root-to-shoot regeneration potential and clonal propagation of Carrizo. Figure S4 Root-to-shoot regeneration potential of different citrus cultivars. Figure S5 Molecular diagnostics of transgenic citrus derived from hairy roots. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.